Comparisons of Defective HTLV-I Proviruses Predict the Mode of Origin and Coding Potential of Internally Deleted Genomes

AbstractCell lines infected with a variety of HTLV-I isolates were examined for the presence of defective proviruses that contain deletions spanning the gag, pol, and env genes. Internally deleted proviruses were identified by Southern blotting and by PCR amplification with 5′ and 3′ primers complementary to gag and tax sequences, respectively. PCR products representing eight defective proviruses from seven different cell lines were subsequently cloned and sequenced. The objectives of this study were twofold: first, we sought to determine whether nucleotide sequences surrounding sites of deletion shared common features that might reveal the mechanisms by which the defective genomes originated. Second, we asked whether deleted proviruses encode Gag fusion proteins with related C-terminal residues derived from open reading frames in the pX region. While most of the defective proviruses had incurred a single, large deletion, two of them displayed a more complex pattern of multiple rearrangements. Alignments of bases flanking the 5′ and 3′ deletion endpoints within each provirus showed tracts of sequence identity consistent with a mechanism involving aberrant intramolecular strand-transfer events during replication. We suggest that the amount or activity of HTLV-I polymerase in virions may contribute both to the poor infectivity of the virus and to the high deletion frequency. Two of the eight proviruses that were examined encoded a gag gene joined to an extended open reading frame; the other six had very short open reading frames (one to six amino acids) derived from pX or env regions joined to gag that showed no apparent amino acid sequence similarity